Asus K8V Delux broken acpi tables?

Asus K8V Delux broken acpi tables?

[Phillip Susi]

As requested, I have attached the output of lspci, dmidecode, dmesg, and 
acpidump.   To keep the size down, they are in a tarball. 

There seems to be something wrong with the acpi tables in the bios 
because when I try to suspend or hibernate I see things like this:

[   32.830712]     ACPI-0508: *** Error: Method execution failed 
[\_SB_.PCI0.IDE C.RATA] (Node ffff81003f0da980), AE_AML_OPERAND_VALUE
[   32.830794]     ACPI-0508: *** Error: Method execution failed 
[\_SB_.PCI0.IDE C.CHN0.DRV0._GTF] (Node ffff81003f0d9100), 
AE_AML_OPERAND_VALUE

Hibernation seems to work anyhow, but suspend to ram does not.  When I 
wake back up from being suspended to ram, the sata disk driver 
complained that a register was in an incorrect state, and the system hung. 

Could someone help me figure out how to correct this?




-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl

Re: Asus K8V Delux broken acpi tables?

[Phillip Susi]

[-- Attachment #1: Type: text/plain, Size: 1342 bytes --]

Oops!  NOW the tarball is attached. 

Phillip Susi wrote:

> As requested, I have attached the output of lspci, dmidecode, dmesg, 
> and acpidump.   To keep the size down, they are in a tarball.
> There seems to be something wrong with the acpi tables in the bios 
> because when I try to suspend or hibernate I see things like this:
>
> [   32.830712]     ACPI-0508: *** Error: Method execution failed 
> [\_SB_.PCI0.IDE C.RATA] (Node ffff81003f0da980), AE_AML_OPERAND_VALUE
> [   32.830794]     ACPI-0508: *** Error: Method execution failed 
> [\_SB_.PCI0.IDE C.CHN0.DRV0._GTF] (Node ffff81003f0d9100), 
> AE_AML_OPERAND_VALUE
>
> Hibernation seems to work anyhow, but suspend to ram does not.  When I 
> wake back up from being suspended to ram, the sata disk driver 
> complained that a register was in an incorrect state, and the system 
> hung.
> Could someone help me figure out how to correct this?
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Power Architecture Resource Center: Free content, downloads, discussions,
> and more. http://solutions.newsforge.com/ibmarch.tmpl
> _______________________________________________
> Acpi-devel mailing list
> Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/acpi-devel
>


[-- Attachment #2: info.tar.gz --]
[-- Type: application/x-gzip, Size: 29644 bytes --]

Re: Asus K8V Delux broken acpi tables?

[Thomas Renninger]

[-- Attachment #1: Type: text/plain, Size: 1617 bytes --]

Phillip Susi wrote:
> Oops!  NOW the tarball is attached.
> Phillip Susi wrote:
> 
>> As requested, I have attached the output of lspci, dmidecode, dmesg, 
>> and acpidump.   To keep the size down, they are in a tarball.
>> There seems to be something wrong with the acpi tables in the bios 
>> because when I try to suspend or hibernate I see things like this:
>>
>> [   32.830712]     ACPI-0508: *** Error: Method execution failed 
>> [\_SB_.PCI0.IDE C.RATA] (Node ffff81003f0da980), AE_AML_OPERAND_VALUE
>> [   32.830794]     ACPI-0508: *** Error: Method execution failed 
>> [\_SB_.PCI0.IDE C.CHN0.DRV0._GTF] (Node ffff81003f0d9100), 
>> AE_AML_OPERAND_VALUE
>>
>> Hibernation seems to work anyhow, but suspend to ram does not.  When I 
>> wake back up from being suspended to ram, the sata disk driver 
>> complained that a register was in an incorrect state, and the system 
>> hung.
>> Could someone help me figure out how to correct this?
>>
I guess that the
         Name (ATA0, Buffer (0x1D) {})
declaration needs to be before it is actually used.

I wonder why the SATA drive only does not work after STR?!?
Attached are my changes, which hopefully let above errors vanish.
Chances are high that this is related to your machine's uncorrect SATA register 
state complains.
I also exchanged some WordAcc to AnyAcc field declarations so that the compiler 
does not complain about wrong field widths, maybe it's even this what causes 
incorrect register reads/writes.

Do you know how to recompile and override your DSDT?

Do you already have the lates BIOS installed? Better check that first.

       Thomas


[-- Attachment #2: dsdt_fixed.diff --]
[-- Type: text/x-patch, Size: 3906 bytes --]

--- dsdt.dsl.orig	2005-10-10 11:19:06.655976980 +0200
+++ dsdt.dsl	2005-10-10 11:34:05.224968654 +0200
@@ -2246,6 +2246,13 @@
                 Name (GSUE, 0x00)
                 Name (GSUT, 0x00)
                 Name (GSCR, 0x00)
+
+		Name (ATA0, Buffer (0x1D) {})
+                Name (ATA1, Buffer (0x1D) {})
+                Name (ATA2, Buffer (0x1D) {})
+                Name (ATA3, Buffer (0x1D) {})
+                Name (ATAB, Buffer (0x1D) {})
+                
                 Device (CHN0)
                 {
                     Name (_ADR, 0x00)
@@ -2473,11 +2480,6 @@
                 {
                     0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x91
                 })
-                Name (ATA0, Buffer (0x1D) {})
-                Name (ATA1, Buffer (0x1D) {})
-                Name (ATA2, Buffer (0x1D) {})
-                Name (ATA3, Buffer (0x1D) {})
-                Name (ATAB, Buffer (0x1D) {})
                 CreateByteField (ATAB, 0x00, CMDC)
                 Method (GTFB, 3, Serialized)
                 {
@@ -2612,19 +2614,19 @@
             {
                 Name (_ADR, 0x00100000)
                 OperationRegion (WAU1, PCI_Config, 0x84, 0x01)
-                Field (WAU1, WordAcc, NoLock, Preserve)
+                Field (WAU1, ByteAcc, NoLock, Preserve)
                 {
                     U184,   8
                 }
 
                 OperationRegion (U1D3, PCI_Config, 0x49, 0x01)
-                Field (U1D3, WordAcc, NoLock, Preserve)
+                Field (U1D3, AnyAcc, NoLock, Preserve)
                 {
                     UR49,   3
                 }
 
                 OperationRegion (UBP1, PCI_Config, 0x85, 0x01)
-                Field (UBP1, WordAcc, NoLock, Preserve)
+                Field (UBP1, AnyAcc, NoLock, Preserve)
                 {
                     U1PE,   1
                 }
@@ -2651,13 +2653,13 @@
             {
                 Name (_ADR, 0x00100001)
                 OperationRegion (WAU2, PCI_Config, 0x84, 0x01)
-                Field (WAU2, WordAcc, NoLock, Preserve)
+                Field (WAU2, AnyAcc, NoLock, Preserve)
                 {
                     U284,   8
                 }
 
                 OperationRegion (UBP2, PCI_Config, 0x85, 0x01)
-                Field (UBP2, WordAcc, NoLock, Preserve)
+                Field (UBP2, AnyAcc, NoLock, Preserve)
                 {
                     U2PE,   1
                 }
@@ -2684,13 +2686,13 @@
             {
                 Name (_ADR, 0x00100002)
                 OperationRegion (WAU3, PCI_Config, 0x84, 0x01)
-                Field (WAU3, WordAcc, NoLock, Preserve)
+                Field (WAU3, AnyAcc, NoLock, Preserve)
                 {
                     U384,   8
                 }
 
                 OperationRegion (UBP3, PCI_Config, 0x85, 0x01)
-                Field (UBP3, WordAcc, NoLock, Preserve)
+                Field (UBP3, AnyAcc, NoLock, Preserve)
                 {
                     U3PE,   1
                 }
@@ -2717,13 +2719,13 @@
             {
                 Name (_ADR, 0x00100003)
                 OperationRegion (WAU4, PCI_Config, 0x84, 0x01)
-                Field (WAU4, WordAcc, NoLock, Preserve)
+                Field (WAU4, AnyAcc, NoLock, Preserve)
                 {
                     U484,   8
                 }
 
                 OperationRegion (UBP4, PCI_Config, 0x85, 0x01)
-                Field (UBP4, WordAcc, NoLock, Preserve)
+                Field (UBP4, AnyAcc, NoLock, Preserve)
                 {
                     U4PE,   1
                 }
@@ -2750,7 +2752,7 @@
             {
                 Name (_ADR, 0x00100004)
                 OperationRegion (WAU5, PCI_Config, 0x84, 0x01)
-                Field (WAU5, WordAcc, NoLock, Preserve)
+                Field (WAU5, AnyAcc, NoLock, Preserve)
                 {
                     U584,   8
                 }

Re: Asus K8V Delux broken acpi tables?

[Phillip Susi]

I think I updated my bios a month or two ago.  I'm not quite sure how to 
recompile my DSDT.  I thought I needed to use iasl to decompile the 
existing DSDT from acpidump, then apply your patch, then recompile.  I 
seem to be missing something though, iasl -d doesn't seem to like my 
acpidump.log. 

Also getting iasl compiled was kind of a pain, the code is a bit 
borked.  First I had to add the -l flag to flex so that unput() could be 
used correctly, then the code was declaring a variable in one place as a 
char [] and in another as a char *, so I had to fix that to get it to 
compile.  Even then it had a bunch of warnings because of incorrect 
casts due to assumptions about being on a 32bit architecture ( I'm 
running 64 bit ). 

Thomas Renninger wrote:

> I guess that the
>         Name (ATA0, Buffer (0x1D) {})
> declaration needs to be before it is actually used.
>
> I wonder why the SATA drive only does not work after STR?!?
> Attached are my changes, which hopefully let above errors vanish.
> Chances are high that this is related to your machine's uncorrect SATA 
> register state complains.
> I also exchanged some WordAcc to AnyAcc field declarations so that the 
> compiler does not complain about wrong field widths, maybe it's even 
> this what causes incorrect register reads/writes.
>
> Do you know how to recompile and override your DSDT?
>
> Do you already have the lates BIOS installed? Better check that first.
>
>       Thomas
>



-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl

Re: Asus K8V Delux broken acpi tables?

[Phillip Susi]

I got iasl -d to disassemble /proc/acpi/dsdt, and your patch cleanly 
applied to that.  When I try to recompile the patched file, I get these 
errors though:

dsdt.dsl    54:             Store (0xAA, DBG8)
Error    1061 -       Object does not exist ^  (DBG8)

dsdt.dsl    58:             Store (0xAC, DBG8)
Error    1061 -       Object does not exist ^  (DBG8)

dsdt.dsl    65:     Method (OSFL, 0, NotSerialized)
Warning  2085 -                ^ Not all control paths return a value (OSFL)

dsdt.dsl    74:             Store (0xAC, DBG8)
Error    1061 -       Object does not exist ^  (DBG8)

dsdt.dsl    78:         If (MCTH (\_OS, "Microsoft Windows NT"))
Error    1094 -                      ^ syntax error, unexpected 
PARSEOP_NAMESEG, expecting ')'

dsdt.dsl    82:         Else
Error    1094 -            ^ syntax error, unexpected PARSEOP_ELSE

dsdt.dsl    84:             If (MCTH (\_OS, "Microsoft WindowsME: 
Millennium Edition"))
Error    1094 -                          ^ syntax error, unexpected 
PARSEOP_NAMESEG, expecting ')'
dsdt.dsl    88:         }
Error    1094 -         ^ syntax error, unexpected '}', expecting $end

Any ideas?


Phillip Susi wrote:

> I think I updated my bios a month or two ago.  I'm not quite sure how 
> to recompile my DSDT.  I thought I needed to use iasl to decompile the 
> existing DSDT from acpidump, then apply your patch, then recompile.  I 
> seem to be missing something though, iasl -d doesn't seem to like my 
> acpidump.log.
> Also getting iasl compiled was kind of a pain, the code is a bit 
> borked.  First I had to add the -l flag to flex so that unput() could 
> be used correctly, then the code was declaring a variable in one place 
> as a char [] and in another as a char *, so I had to fix that to get 
> it to compile.  Even then it had a bunch of warnings because of 
> incorrect casts due to assumptions about being on a 32bit architecture 
> ( I'm running 64 bit ).
>


-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl

Re: Asus K8V Delux broken acpi tables?

[Thomas Renninger]

Phillip Susi wrote:
> I think I updated my bios a month or two ago.  I'm not quite sure how to 
> recompile my DSDT.  I thought I needed to use iasl to decompile the 
> existing DSDT from acpidump, then apply your patch, then recompile.  I 
> seem to be missing something though, iasl -d doesn't seem to like my 
> acpidump.log.
> Also getting iasl compiled was kind of a pain, the code is a bit 
> borked.  First I had to add the -l flag to flex so that unput() could be 
> used correctly, then the code was declaring a variable in one place as a 
> char [] and in another as a char *, so I had to fix that to get it to 
> compile.  Even then it had a bunch of warnings because of incorrect 
> casts due to assumptions about being on a 32bit architecture ( I'm 
> running 64 bit ).

I used ACPICA 20050902 compiler.
Hopefully there is nothing broken in newer versions?
I will give it a try if I find the time ...
I will send you the compiled versions privately.

       Thomas


-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl

Re: Asus K8V Delux broken acpi tables?

[Stefan Seyfried]

On Sat, Oct 08, 2005 at 01:44:59AM -0400, Phillip Susi wrote:
 
> Hibernation seems to work anyhow, but suspend to ram does not.  When I 
> wake back up from being suspended to ram, the sata disk driver 
> complained that a register was in an incorrect state, and the system hung. 

Which kernel version?
The ACPI error and the suspend to RAM failure could be entirely different
problems, SATA in mainline kernel has only very recently got suspend/resume
support IIRC.
-- 
Stefan Seyfried



-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl